Motor Shaft Vibration Signal Detection Based on Monocular Visio

To address the limitations of traditional vibration sensors in measuring rotational body displacement and the high costs of eddy current sensors and laser Doppler vibrometers, this paper proposes a novel monocular camera-based method for motor shaft vibration displacement measurement, integrating the CSRT tracker and Hough circle detection algorithm. First, the method establishes pixel-to-camera coordinate correspondence through camera calibration and maps pixels to camera coordinates using depth information calculation. Then, an industrial camera captures rotor vibration videos, where the CSRT tracker locates the motor shaft position. And then, the geometric center is precisely refined via Hough circle detection to correct the tracking center, resolving the offset issue of traditional tracking bounding boxes. At last, under known camera intrinsics and depth information, pixel coordinates are mapped to camera's intrinsic parameters to derive the vibration amplitude curve of the motor shaft. Experimental results under typical operating conditions demonstrate the method's accuracy and stability. By incorporating Hough circle detection, the average error is significantly reduced to <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{1 - 2}$</tex> pixels compared to conventional CSRT algorithms, with tracking errors remaining below 2 pixels (equivalent to approximately <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{0. 3 ~ m m}$</tex> in physical distance) at <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{1 3 5 0}$</tex> RPM. This non-contact approach is suitable for motor health monitoring and fault diagnosis, showing potential applications in aerospace, power generation equipment, intelligent manufacturing, and other fields for equipment condition assessment. assessment.